Creating a predictive measure

Last modified by Aurelie Bertrand on 2026/07/01 11:50

You can create predictive measures based on measures and calculated measures. These measures enable you to predict the value of a measure based on a time dimension.

To create a predictive measure:

  1. Click the New Measure button, then, in the context menu that appears, select Create Predictive Measure.
    New measure
    ➡ The Predictive Measure dialog box appears.
    Predictive measure
  2. Enter the Name you wish to give your predictive measure, then select the Category.
  3. In the Columns section, specify:
    • The Temporal dimension according to which you wish to explore/predict this measure
    • The Measure to be predicted
    • The Aggregation of this measure
  4. In the Model type section , enter :
    • The algorithm to be used to predict the measure selected in the Columns section.
    • If the selected algorithm is the moving average, you must also specify the Number of data points over which the moving average is calculated (order, default value is 2).
    • The value to be predicted (measure value, lower or upper bound of the confidence interval).

💡 See the page Predictive Measures for more details on choosing a model.

  1. You can extend the time dimension you have selected in the Column group by clicking the corresponding button. For more information on how to add members to the date, please refer to the section Adding data to a dimension.

💡 For each algorithm, we display a brief description (conditions of use) next to the help icon in the bottom left-hand corner. In addition, you can click the ‘Help model choice’ box to receive guidance on choosing an algorithm. This opens the ‘Help model choice’ window:

Help

By answering the two questions shown in the image above, we narrow down the list of algorithms to make your choice easier.